Skip to contents

Test Proportional Hazards Assumption

Source: R/efficacy_tte.R
test_ph_assumption.Rd

Tests the proportional hazards assumption for Cox regression models using Schoenfeld residuals (cox.zph test).

Usage

test_ph_assumption(
  data,
  time_var = NULL,
  event_var = NULL,
  trt_var = "TRT01P",
  covariates = character(),
  alpha = 0.05,
  plot = FALSE
)

Arguments

data

An ADaMData object (with domain "ADTTE") or an ADTTE data frame with time-to-event data, or a coxph model object

time_var

Character. Time variable (required if data is a data frame)

event_var

Character. Event variable (required if data is a data frame)

trt_var

Treatment variable name (default: "TRT01P"). Ignored for ADaMData objects which use their own trt_var property.

covariates

Character vector. Additional covariates to include

alpha

Numeric. Significance level for flagging violations (default: 0.05)

plot

Logical. Whether to create diagnostic plot (default: FALSE)

Value

A list with:

  • results: Data frame with variable, rho, chisq, p-value, violation flag

  • global_test: Global test result (p-value)

  • violation: Logical, TRUE if any p < alpha

  • model: The fitted coxph model

  • zph: The cox.zph result object

  • plot: ClinicalPlot if plot=TRUE

References

Grambsch, P. M. and Therneau, T. M. (1994). Proportional hazards tests and diagnostics based on weighted residuals. Biometrika, 81, 515-26.

IQWiG Methods v8.0, Section 10.3.12, p. 235-237.

Examples

if (FALSE) { # \dontrun{
result <- test_ph_assumption(
  data = adtte,
  time_var = "AVAL",
  event_var = "CNSR",
  trt_var = "TRT01P",
  plot = TRUE
)
} # }